vLLM test reports 13% higher Step5 throughput with PrimTS on Blackwell
The merged, opt-in FlashInfer path reported 147.04 tokens per second at roughly 96K input and concurrency 32, and 5.438-second time to first token for a separate 160K batch-one case on four GB300s.
By RuntimeWire Staff · Published

The merged, opt-in FlashInfer path reported 147.04 tokens per second at roughly 96K input and concurrency 32, and 5.438-second time to first token for a separate 160K batch-one case on four GB300s.